Skip to main content
NCBI home page
Science Advances logoLink to Science Advances
. 2024 Mar 22;10(12):eadk5177. doi: 10.1126/sciadv.adk5177

Bone-inspired stress-gaining elastomer enabled by dynamic molecular locking

Yang Wang 1,†, Qingbao Guan 1,†, Yue Guo 1, Lijie Sun 1, Rasoul Esmaeely Neisiany 2,3, Xuran Guo 4, Hongfei Huang 1, Lei Yang 1, Zhengwei You 1,*
PMCID: PMC10959417  PMID: 38517961

Abstract

The limited capacity of typical materials to resist stress loading, which affects their mechanical performance, is one of the most formidable challenges in materials science. Here, we propose a bone-inspired stress-gaining concept of converting typically destructive stress into a favorable factor to substantially enhance the mechanical properties of elastomers. The concept was realized by a molecular design of dynamic poly(oxime-urethanes) network with mesophase domains. During external loading, the mesophase domains in the condensed state were aligned into more ordered domains, and the dynamic oxime-urethane bonds served as the dynamic molecular locks disassociating and reorganizing to facilitate and fix the mesophase domains. Consequently, the tensile modulus and strength were enhanced by 1744 and 49.3 times after four cycles of mechanical training, respectively. This study creates a molecular concept with stress-gaining properties induced by repeated mechanical stress loading and will inspire a series of innovative materials for diverse applications.

The destructive stress has been converted into a favorable factor to substantially enhance the material mechanical properties.

INTRODUCTION

Materials dominate all aspects of our life to such an extent that it is difficult to imagine life without them (1–4). Mechanical loading environments and internal stress concentration cause great damage to the typical materials (5, 6) and remain a critical concern in the field of materials science. The internal stress gained during processing affects the mechanical stability of materials and dramatically shortens their durability and reliability. Furthermore, external stresses are repeatedly loaded on materials during service life, which accelerates aging and may eventually lead to their failure.

The stress loading effect (SLE; strength changing after stress loading and unloading divided by the original tensile strength) is proposed to quantify the change in tensile strength. The SLEs of traditional polymers are negative. As a result of the disentanglements and slippages of molecular chains, high-density polyethylene (7) and thermoplastic polyurethane (8), both of which are thermoplastic (linear) polymers, exhibited an SLE of −0.0905 and −0.2441, respectively. In the case of thermosetting (cross-linked) polymers, the vulcanized natural rubber (9) also exhibited a negative SLE of −0.5506. Although their slippage during stretching is substantially retarded by the cross-linked network, which makes macromolecules difficult to disentangle, the restrained slippage hinders the internal stress to relax completely, and the accumulated internal stress causes the fracture of molecular chains until the catastrophic mechanical failure.

In contrast to the negative SLEs observed in traditional polymers, several polymers with positive SLEs have been reported recently (10–15). These strategies are classified into two types. The first strategy is a physical rearrangement. Lin et al. (10) propose a mechanical training strategy to form aligned nanofibrils for the development of fatigue-resistant hydrogels. Inspired by the dynamic sacrificial bonds in biomaterials and the self-strengthening mechanism of skeletal muscles, Tu et al. (11) fabricated programmable artificial muscle material. Agrawal et al. (12) developed dynamic self-stiffening in liquid crystal elastomers under repeated low-strain compressive deformation. Liu et al. (15) used a strain-induced crystallization strategy to prepare a nondestructively reinforced hydrogel. However, in the absence of a chemical cross-linking network to fix the physical rearrangement, it is difficult to ensure its stability upon heating and during use. The second strategy is covalent network growth. By stabilizing the input of monomers in chemical reactions, Matsuda et al. (13) developed self-growing hydrogels induced by mechanophores. However, the requirement of continuous monomer supplements is usually infeasible in practice. Seshimo et al. (14) prepared self-strengthening polyurethane elastomers by force-induced cross-linking reactions. The cross-linking reaction of the pre-reserved reactive groups was used to enhance the storage modulus; however, it embrittled the polymer. Yu et al. (16) used a water-assisted cross-linking reaction mechanism to prepare a self-strengthening polymer, which, however, required aquatic environment. In addition, both SLE and modulus changes after stress loading for all the developed polymers were limited.

Bone is regarded as a natural stress-gaining material (17–19). The bone structure is capable of remolding to strengthen itself upon cyclic training with appropriate stress. External tension induces dissociation (unlocking) of the covalent cross-links between collagens. The consequent unlocked mineralized collagen fibrils are aligned into highly ordered structures and covalently cross-linked (locked) again upon unloading to substantially improve the stiffness and strength of the bone (Fig. 1A) (20, 21). Here, we propose a bone-inspired “dynamic molecular locking” strategy to use the loading stress for polymer strengthening. A covalently cross-linked polyurethane with mesophase domains (MCPU) was designed using the reversible oxime-urethane bonds as dynamic molecular locks, the dissociation of which allowed a gradual release of stress. Meanwhile, the stress induced the rigid segments to align into highly ordered domains in the condensed state, which were locked in the reorganized polymeric covalent network based on dynamic oxime-urethane bonds (see “Design, characterization of polymers and the optimization of training process” in the Supplementary Materials). The cyclic alignment and locking of mesophase domains continuously increased the molecular order to substantially strengthen and toughen the resultant MCPU polymers (Fig. 1, A and B).

Fig. 1. Design and outcome of stress-gaining polymers.

Fig. 1.

Open in a new tab

(A) Schematics of the stress-gaining process of native bone (a) and covalently cross-linked polymers via dynamic molecular locking (b). (B) Strength and modulus changes after stress loading of traditional, state-of-the-art self-strengthening polymers and our polymer [(9): traditional materials; (15): slide-ring hydrogels; (12): polydomain nematic liquid crystal elastomers; (10): freeze-thawed polyvinyl alcohol hydrogel; (14): segmented polyurethane elastomers; (11): poly(ethylene-propylene-diene monomer) elastomer; (13): double-network hydrogels; (16): water-strengthening polyurethane].

RESULTS

Synthesis of molecules

Dimethylglyoxime (DMG)–based cross-linked polyurethane with mesophase domains (DMG-MCPU) was prepared through a facile one-pot condensation polymerization from commercially available benzoic acid 4-hydroxy-4-hydroxyphenyl ester, polyethylene glycol (PEG) 400, 4,4-diphenylmethane diisocyanate (MDI), DMG, and glycerol, in the presence of the catalyst dibutyltin dilaurate (DBTDL) (fig. S1). The synthesized polymer structures were characterized by attenuated total reflectance Fourier transform infrared (ATR-FTIR) (fig. S2). Swelling experiments were also carried out to confirm the cross-linked structures of both prepared MCPUs (fig. S3).

Structural evolution under stress

To investigate the locking role of the dynamic oxime-urethane bonds in stress gaining, the DMG monomer was replaced with 1,4-butanediol (BDO) to prepare a reference polymer (BDO-MCPU), which is a traditional covalently cross-linked polyurethane (fig. S1). As designed, DMG-MCPU was expected to reorganize the network at room temperature, and the reference BDO-MCPU was expected to exhibit poor mobility, although it might also be induced by transcarbamoylation at high temperature (Fig. 2A) (22–24). In situ variable temperature FTIR spectroscopy revealed the thermal reversibility of the dynamic urethane-oxime bonds, which was essential for the reorganization of the covalent network (Fig. 2B). The -N═C═O group corresponding peak of 2271 cm−1 appeared in DMG-MCPU sample upon heating and became stronger with the increase in temperature, indicating the dissociation of the oxime-urethane bonds. However, no apparent -N═C═O peak was observed in the FTIR spectrum of BDO-MCPU at the same conditions, indicating its stability. Both MCPU exhibited a similar trend in storage modulus (Fig. 2C), with a rubber plateau above Tg (~20°C) (fig. S4) Owing to the dynamic association and dissociation of the reversible oxime-urethane bonds, DMG-MCPU exhibited a lower storage modulus than BDO-MCPU over the entire measurement temperature range (−20°C to 150°C). The relaxation time of DMG-MCPU for a 150% strain at 95°C was determined to be 56 min, while the relaxation time of control sample BDO-MCPU was more than 380 min under the same condition (Fig. 2D). These results demonstrated that urethane oxime bond–based dynamic molecular lock effectively promoted the covalent network reorganization.

Fig. 2. The schematic, characterization, and outcome of the reorganization of the synthesized MCPU network.

Fig. 2.

Open in a new tab

(A) Structures and network evolution of DMG-MCPU with dynamic oxime-polyurethane molecular lock and reference BDO-MCPU. (B) In situ variable temperature FTIR spectra of DMG-MCPU and BDO-MCPU. (C) Storage modulus (E’) of DMG-MCPU and BDO-MCPU as a function of temperature. (D) Stress relaxation curves of the synthesized MCPUs. (E) Dimension and strength retention ratio after heating of trained (single cyclic) DMG-MCPU and BDO-MCPU strips. (F) Comparison of dimensional stability of DMG-MCPU and BDO-MCPU strips. (a) Original strips. (b) Trained (single cyclic) strips. (c) Strip b heated at 140°C for 20 min (scale bar, 10 mm).

Furthermore, the stability of the reorganized network of the trained DMG-MCPU was essential for its long-term performance and was evaluated by the dimensional and mechanical stability. The DMG-MCPU strip was trained by the procedure as shown in fig. S5, and we selected 95°C as the training temperature by stress relaxation experiment of DMG-MCPU (fig. S6). The trained DMG-MCPU strip well-kept the length, while the trained BDO-MCPU strip shrank by ~25% under the same conditions (Fig. 2, E and F, and fig. S7). The tensile stress at 50% strain of the trained MCPUs before and after heating (210°C for 10 s) was compared. The stress retention ratio (stress after heating divided by stress before heating) in the trained DMG-MCPU was 96 ± 4%, whereas that in the BDO-MCPU was only 76 ± 6% (Fig. 2E and fig. S8). These results confirmed the efficient reorganization and relocking of the DMG-MCPU network, where “dynamic molecular locking” played a key role. Thermogravimetric analysis (TGA) results revealed that the MCPU was stable up to 210°C (fig. S9). The glass transition of DMG-MCPU appeared at 26.5°C and increased to 35.6°C after four cycles of training (fig. S4). Polarized optical microscope (POM) monitored that the birefringence of trained fourth DMG-MCPU film remained stable up to 200°C (fig. S10), which confirmed the high mobility of the molecular chains to facilitate the reorganization of dynamic oxime-based polyurethane network and guarantee cumulatively steady alignment of mesophase domains.

The orientation of the rigid segments under tension is critical for the stress gaining of the synthesized MCPU (Fig. 3A) (12). The evolution of mesophase domains in the condensed state of the DMG-MCPU strip before and after one (Tr1st), two (Tr2nd), three (Tr3rd), and four times (Tr4th) of training was characterized by POM. The threaded texture and speckles were scattered in the as-prepared DMG-MCPU strip. The micro-domains were small and isolated within an isotropic PU matrix. Much more paralleled texture formed in the trained sample (Fig. 3B), and a pair of gradual enhanced diffraction arcs appeared at the meridian (Fig. 3C). These results implied substantially improved alignment and agglomeration of molecular chains. It is convinced by the fact that DMG-MCPU with dynamic molecular locks exhibited a larger increasement (up to four times) in order parameter (Hermans orientation factor) than that of BDO-MCPU and their linear counterpart (LCPU) under the same training times (Fig. 3D and figs. S11 to S17). The intensity of DMG-MCPU, BDO-MCPU, and LCPU was plotted as a function of q-vector based on the two-dimensional wide-angle x-ray scattering (2D WAXD) data (fig. S18). The d-spacing of all these materials decreased from ~4.4 Å to ~4.3 Å. This indicates that the distance between molecular chains reduces with increased stretching, resulting in a more orderly mesophase orientation. Atomic force microscope (AFM) (Fig. 3E) images revealed that the modulus of the Tr1st DMG-MCPU strip was one order magnitude higher than that of the as-prepared DMG-MCPU strip. Meanwhile, an increase in the rigid phase was observed due to the agglomeration of mesophase domains.

Fig. 3. The orientation during the reorganization of DMG-MCPU with dynamic molecular lock.

Fig. 3.

Open in a new tab

(A) Schematic of the orientation of rigid segment of DMG-MCPU. (B) POM images of DMG-MCPU films. (a) Disordered texture of the original film. The threaded texture of Tr2nd (b) and Tr4th (c) samples (the arrows denoted the stretching direction). (C) 2D WAXD patterns of the original, Tr2nd, and Tr4th DMG-MCPU films at room temperature (the x-ray beam is perpendicular to the film direction). (D) Enhanced alignment of DMG-MCPU, BDO-MCPU, and LCPU after training different times. a.u., arbitrary units. (E) AFM images of the original and Tr1st DMG-MCPU and their corresponding modulus.

Evaluation of the stress-gaining behavior

The tensile tests of DMG-MCPU strips were carried out to further evaluate the mechanical properties and stress-gaining behavior of the synthesized MCPUs during cyclic training. The original DMG-MCPU and BDO-MCPU samples were both comparable to those of linear counterpart (LCPU) and typical liquid crystal elastomers in terms of elastic modulus (0.72 ± 0.17 and 0.45 ± 0.24 MPa), tensile strength (0.66 ± 0.02 and 0.76 ± 0.01 MPa), and toughness (4.93 ± 0.40 and 5.63 ± 0.35 J/m3) (Fig. 4A and tables S1 and S2) (25–27). However, MCPU exhibited a mechanical enhancement that was substantially distinct from LCPU after the first training. The tensile strength of BDO-MCPU and DMG-MCPU increased by 10.8 and 21.2 times, respectively, while only a limited improvement (1.21 times) occurred in LCPU (table S3). After three times training, the elastic modulus (504.46 ± 33.81 MPa and 95.8 ± 19.61 MPa), tensile strength (41.66 ± 2.51 MPa and 33.45 ± 0.37 MPa), and toughness (260 ± 20.22 J/m3 and 81.35 ± 12.31 J/m3) of DMG-MCPU and BDO-MCPU increased. BDO-MCPU fractured after four training cycles, while DMG-MCPU exhibited continued improvement in mechanical properties. The mechanical enhancement behavior in MCPU is similar to bone strengthening through external tension, so-called “stress gaining.”

Fig. 4. Boosting of the mechanical properties of stress-gaining DMG-MCPU after training.

Fig. 4.

Open in a new tab

(A) Tensile stress-strain curves of MCPUs and uncross-linked reference LCPU strips after 1, 2, 3, and 4 times of training. (B) Schematic of microstructural evolution of Tr3rd DMG-MCPU during stretching. Stress-strain curves at different stretching rates for the Tr3rd (C) BDO-MCPU and (D) DMG-MCPU. Comparison of true strengths (E) and elastic moduli (F) of the state-of-the-art self-strengthening materials and DMG-MCPU for being trained at different times.

With the physical alignment and enlarged mesophase domains during training, the dynamic oxime-urethane bonds allowed for the reorganization of the covalent networks to adapt to the tension and lock the orientation, resulting in improved strength and toughness for DMG-MCPU. The weak intermolecular interactions of LCPU make it prone to chain slippage when subjected to external forces, which are not conducive to the formation and orientation of mesophase domains, limiting the improvement of strength. As a representative example, the schematic of the microstructural evolution of Tr3rd DMG-MCPU during the tensile test is shown in Fig. 4B. At the initial stage of deformation (a), DMG-MCPU showed a sharp linear tensile stress-strain curve revealing a high elastic modulus, likely being induced by the high orientation during training (fig. S19). As the strain increased (b), separation and slippage of mesophase domains occurred, and necking was observed. Furthermore, the fragmentation of mesophase domains occurred in stage (c), accompanied by the unfolding and reorganization of molecular chains. The energy dissipation of dynamic oxime-urethane bonds contributed to greater deformation. With the continuous reorganization of the covalent polymeric network and alignment of rigid domains in the last strain hardening stage (d), the formation of numerous microfibrils structure was observed (fig. S20) (28), resulting in a large strain and high strength. It was observed that DMG-MCPU and BDO-MCPU had distinct sensitivities to tensile rate (Fig. 4, C and D, and fig. S21). Under the same tensile test condition (20 mm/min), the trained DMG-MCPU strip with oxime-urethane bonds exhibited a typical ductile fracture in stress-strain curves rather than a brittle fracture of BDO-MCPU. It suggested that the presence of dynamic molecular locks in DMG-MCPU suppresses local stress amplification more effectively under a high stretching rate, providing effective stress dissipation and reducing the damage caused by crazing and stretching segments (29).

Eventually, the mechanical properties of DMG-MCPU were greatly improved (Fig. 4, A, E, and F, and table S1). In contrast, uncross-linked LCPU showed limited mechanical improvement (3.7, 5.5, and 2.1 times) after the same training procedure. The dynamic molecular locking in dynamic covalent network and external loading induced the reorganization of mesophase domains synergistically and enabled a substantial stress-gaining effect, which greatly strengthened the DMG-MCPU strips, superior to traditional materials with weakening in the mechanical properties that responded to external stress. The tensile strength of DMG-MCPU was increased by 49.3 times after simple four times training, which substantially exceeded that of previously reported self-strengthening materials (SLE < 1.5) (Fig. 4E). The modulus was also enhanced by 1744 times, which was unprecedented comparing the state-of-the-art self-strengthening polymers (Fig. 4F).

Potential applications of the developed stress-gaining materials

The stress-gaining effect of DMG-MCPU can be used to construct adaptively strengthening materials to match dynamic requirements such as an adaptive ligament wrap bandage (Fig. 5A). Original DMG-MCPU, in its soft state, serves as a compliant bandage to avoid damage to the injured ligament (figs. S22 and S23). The bandage is becoming stronger during rehabilitation training to well adapt to the increased support requirement of recovering ligament. To demonstrate the feasibility of this application, the in vitro biocompatibility of DMG-MCPU was performed. The proliferation of mouse fibroblast cell line L929 cultured with DMG-MCPU was similar to that of cells cultured with positive control (polycaprolactone) (fig. S24). The mechanical behavior of the DMG-MCPU bandage was evaluated by training. The original DMG-MCPU bandage was soft and well adapt to external loading. Then, its mechanical properties were gradually improved through subsequent loading. As desired, the DMG-MCPU bandage became much stronger and could easily lift 2500-g weight (4900 times its own weight) after training (Fig. 5B). Furthermore, the trained DMG-MCPU bandage exhibited outstanding scratch resistance, which was also beneficial to prevent further injury (Fig. 5C). Furthermore, in addition to the self-gaining of the adaptive ligament bandage, DMG-MCPU also exhibited self-healing and shape memory properties (figs. S25 and S26), with potential applications in various fields, such as minimally invasive implants, mechanical connectors, conveyor belts, and smart soft robots. We believe that this work will inspire a family of innovative concept materials and enable innovative applications.

Fig. 5. The potential application demonstration of the prepared DMG-MCPU.

Fig. 5.

Open in a new tab

(A) Schematic of the application as an adaptive ligament wrapping bandage. (B) Apparent deformation of the original DMG-MCPU after being loaded with 50 g weight, whereas trained DMG-MCPU could hang 2500-g weight. (C) The original DMG-MCPU was vulnerable to scratch. After training, DMG-MCPU showed excellent scratch resistance (scale bar, 10 mm).

DISCUSSION

Here, we proposed a bone-inspired material concept of stress gaining. We proved that a dynamic poly(oxime-urethanes) with mesophase domains converted typical destructive stress into a favorable factor to substantially strengthen itself, distinguishing it from traditional stress-weakening materials subjected to applied stresses. The stress induced the rigid segment to be aligned into ordered mesophase domains in the condensed state, which was facilitated and locked by the reorganization of the covalent polymeric network via the reversible lock and unlock of the oxime-urethane bonds. The responsive stress-gaining effect on the mechanical properties via dynamic molecular locking strategy in our study is unprecedented, with an increase in the elastic modulus and tensile strength by three and one order of magnitude, respectively. The stress-gaining concept represents a paradigm in material design. This will inspire innovative molecular strategies to improve the performance of materials under dynamic mechanical loads.

MATERIALS AND METHODS

Materials

PEG [number-average molecular weight (Mn) = ~400 g mol−1] and DBTDL (95%) were both purchased from Aladdin. MDI (99%) was purchased from Wanhua Chemical Group. DMG (98%) and glycerol (99%) were purchased from Sinopharm Chemical Reagent. 4-Hydroxyphenyl-4-hydroxybenzoate (HPHB) was supplied by Yuhao Chemical. N,N-dimethylformamide (DMF) (99.8%) was supplied by J&K. BDO was purchased from Damas-beta Reagent. All the chemicals and reagents were used as received without further purification unless otherwise noted.

Synthesis of MCPU and LCPU

DMG-MCPU

PEG (3.257 g, 8.142 mmol) was poured into the reactor and dried at 110°C for 120 min in a vacuum. The reactor containing the PEG was then cooled to 40°C. HPHB (1 g, 4.343 mmol) and DMG (0.1260 g, 1.085 mmol) were dissolved in 8 ml of DMF, and then the mixture was added to the cooled reactor containing dried PEG. DBTDL (0.5 wt %, 0.04 g), glycerol (0.0428 g, 0.475 mmol), and MDI (3.575 g, 14.28 mmol) were sequentially added to the mixture. The reaction occurred at 75°C for 8 hours under a nitrogen atmosphere. Afterward, the reaction mixture was poured into a polytetrafluoroethylene (PTFE) mold to volatilize DMF at 30°C for 18 hours. The reaction was allowed to complete by putting the PTFE mold in an oven, while the temperature was increased from 30°C to 75°C for another 48 hours. After complete curing, the resultant mixture was transferred to a vacuum oven and heated at 75°C for another 24 hours.

BDO-MCPU

PEG (3.257 g, 8.142 mmol) was poured into the reactor and dried at 110°C for 120 min in a vacuum. Here, also, the reactor was then cooled to 40°C. HPHB (1 g, 4.343 mmol) and BDO (0.0978 g, 1.085 mmol) were dissolved in 8 ml of DMF, and the mixture was then added to the cooled reactor containing dried PEG. DBTDL (0.5 wt %, 0.04 g), glycerol (0.0428 g, 0.475 mmol), and MDI (3.575 g, 14.28 mmol) were sequentially added to the mixture. The synthesis procedure was the same as that of DMG-MCPU.

LCPU

PEG (3.257 g, 8.142 mmol) was poured into the reactor and dried at 110°C for 120 min in a vacuum. Then, the reactor was cooled to 40°C. HPHB (1 g, 4.343 mmol) and DMG (0.1260 g, 1.085 mmol) were dissolved in 8 ml of DMF, and the mixture was then added to the cooled reactor containing dried PEG. DBTDL (0.5 wt %, 0.04 g), 1,3-propanediol (0.0542 g, 0.712 mmol), and MDI (3.575 g, 14.28 mmol) were sequentially added to the mixture. The synthesis procedure was the same as that of DMG-MCPU.

Acknowledgments

We thank Y. Men and S. Li (Changchun Institute of Applied Chemistry, Chinese Academy of Sciences) for the assistance with x-ray diffraction analysis. We also thank Philippe Poulin (University of Bordeaux) for valuable discussion on the mechanism of stress-induced stress gaining.

Funding: This work was supported by the National Key Research and Development Program of China (2021YFC2101800 and 2021YFC2400802), the National Natural Science Foundation of China (52173117, 52073049, and 21991123), the Shanghai Rising-Star Program (21QA1400200), the Natural Science Foundation of Shanghai (22ZR1400700), the Ningbo 2025 Science and Technology Major Project (2019B10068), and the Science and Technology Commission of Shanghai (20DZ2254900 and 20DZ2270800).

Author contributions: Z.Y., Y.W., and Y.G. conceived the concept and designed the experiments. Z.Y. supervised the whole project. Y.W. and Y.G. performed the experiments. X.G. performed the biocompatibility evaluation. Q.G., L.S., H.H., and L.Y. assisted in data analysis. Y.W., Q.G., R.E.N., and Z.Y. wrote and reviewed the manuscript.

Competing interests: Z.Y., Y.W., and Y.G. are the inventors on a provisional patent application related to this work filed by Donghua University (serial no. 202210514463.4, dated 11 May 2022). The authors declare that they have no other competing interests.

Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials.

Supplementary Materials

This PDF file includes:

Supplementary Text

Figs. S1 to S26

Tables S1 to S3

Legends for movies S1 to S4

References

sciadv.adk5177_sm.pdf (2MB, pdf)

Other Supplementary Material for this manuscript includes the following:

Movies S1 to S4

sciadv.adk5177_movies_s1_to_s4.zip (54.1MB, zip)

REFERENCES AND NOTES

  • 1.Ramirez A. L. B., Kean Z. S., Orlicki J. A., Champhekar M., Elsakr S. M., Krause W. E., Craig S. L., Mechanochemical strengthening of a synthetic polymer in response to typically destructive shear forces. Nat. Chem. 5, 757–761 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Wang Z., Wang J., Ayarza J., Steeves T., Hu Z., Manna S., Esser-Kahn A. P., Bio-inspired mechanically adaptive materials through vibration-induced crosslinking. Nat. Mater. 20, 869–874 (2021). [DOI] [PubMed] [Google Scholar]
  • 3.Mohanty A. K., Wu F., Mincheva R., Hakkarainen M., Raquez J.-M., Mielewski D. F., Narayan R., Netravali A. N., Misra M., Sustainable polymers. Nat. Rev. Methods Primers 2, 46 (2022). [Google Scholar]
  • 4.Guo Z., Lu X., Wang X., Li X., Li J., Sun J., Engineering of chain rigidity and hydrogen bond cross-linking toward ultra-strong, healable, recyclable, and water-resistant elastomers. Adv. Mater. 35, e2300286 (2023). [DOI] [PubMed] [Google Scholar]
  • 5.Chen Y., Yeh C. J., Qi Y., Long R., Creton C., From force-responsive molecules to quantifying and mapping stresses in soft materials. Sci. Adv. 6, eaaz5093 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Sapora A., Cornetti P., Carpinteri A., Firrao D., Brittle materials and stress concentrations: Are they able to withstand? Procedia Eng. 109, 296–302 (2015). [Google Scholar]
  • 7.Hiss R., Hobeika S., Lynn C., Strobl G., Network stretching, slip processes, and fragmentation of crystallites during uniaxial drawing of polyethylene and related copolymers. A comparative study. Macromolecules 32, 4390–4403 (1999). [Google Scholar]
  • 8.Scetta G., Ju J., Selles N., Heuillet P., Ciccotti M., Creton C., Strain induced strengthening of soft thermoplastic polyurethanes under cyclic deformation. J. Polym. Sci. 59, 685–696 (2021). [Google Scholar]
  • 9.Mullins L., Effect of stretching on the properties of rubber. Rubber Chem. Technol. 21, 281–300 (1948). [Google Scholar]
  • 10.Lin S., Liu J., Liu X., Zhao X., Muscle-like fatigue-resistant hydrogels by mechanical training. Proc. Natl. Acad. Sci. U.S.A. 116, 10244–10249 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Tu Z., Liu W., Wang J., Qiu X., Huang J., Li J., Lou H., Biomimetic high performance artificial muscle built on sacrificial coordination network and mechanical training process. Nat. Commun. 12, 2916 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Agrawal A., Chipara A. C., Shamoo Y., Patra P. K., Carey B. J., Ajayan P. M., Chapman W. G., Verduzco R., Dynamic self-stiffening in liquid crystal elastomers. Nat. Commun. 4, 1739 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Matsuda T., Kawakami R., Namba R., Nakajima T., Gong Jian P., Mechanoresponsive self-growing hydrogels inspired by muscle training. Science 363, 504–508 (2019). [DOI] [PubMed] [Google Scholar]
  • 14.Seshimo K., Sakai H., Watabe T., Aoki D., Sugita H., Mikami K., Mao Y., Ishigami A., Nishitsuji S., Kurose T., Ito H., Otsuka H., Segmented polyurethane elastomers with mechanochromic and selfstrengthening functions. Angew. Chem. Int. Ed. Engl. 60, 8406–8409 (2021). [DOI] [PubMed] [Google Scholar]
  • 15.Liu C., Morimoto N., Jiang L., Kawahara S., Noritomi T., Yokoyama H., Mayumi K., Ito K., Tough hydrogels with rapid self-reinforcement. Science 372, 1078–1081 (2021). [DOI] [PubMed] [Google Scholar]
  • 16.Yu K., Feng Z., Du H., Lee K. H., Li K., Zhang Y., Masri S. F., Wang Q., Constructive adaptation of 3D-printable polymers in response to typically destructive aquatic environments. PNAS Nexus 1, pgac139 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Koons G. L., Diba M., Mikos A. G., Materials design for bone-tissue engineering. Nat. Rev. Mater. 5, 584–603 (2020). [Google Scholar]
  • 18.Nudelman F., Kröger R., Enhancing strength in mineralized collagen. Science 376, 137–138 (2022). [DOI] [PubMed] [Google Scholar]
  • 19.Christen P., Ito K., Ellouz R., Boutroy S., Sornay-Rendu E., Chapurlat R. D., van Rietbergen B., Bone remodelling in humans is load-driven but not lazy. Nat. Commun. 5, 4855 (2014). [DOI] [PubMed] [Google Scholar]
  • 20.Ping H., Wagermaier W., Horbelt N., Scoppola E., Li C., Werner P., Fu Z., Fratzl P., Mineralization generates megapascal contractile stresses in collagen fibrils. Science 376, 188–192 (2022). [DOI] [PubMed] [Google Scholar]
  • 21.Gachon E., Mesquida P., Stretching single collagen fibrils reveals nonlinear mechanical behavior. Biophys. J. 118, 1401–1408 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Zheng N., Fang Z., Zou W., Zhao Q., Xie T., Thermoset shape-memory polyurethane with intrinsic plasticity enabled by transcarbamoylation. Angew. Chem. Int. Ed. Engl. 55, 11421–11425 (2016). [DOI] [PubMed] [Google Scholar]
  • 23.Jin B., Yang S., Programming liquid crystalline elastomer networks with dynamic covalent bonds. Adv. Funct. Mater. 33, 2304769 (2023). [Google Scholar]
  • 24.Liang H., Zhang S., Liu Y., Yang Y., Zhang Y., Wu Y., Xu H., Wei Y., Ji Y., Merging the interfaces of different shape-shifting polymers using hybrid exchange reactions. Adv. Mater. 35, e2202462 (2023). [DOI] [PubMed] [Google Scholar]
  • 25.White T. J., Broer D. J., Programmable and adaptive mechanics with liquid crystal polymer networks and elastomers. Nat. Mater. 14, 1087–1098 (2015). [DOI] [PubMed] [Google Scholar]
  • 26.Wang Y., Guan Q., Lei D., Esmaeely Neisiany R., Guo Y., Gu S., You Z., Meniscus-climbing system inspired 3D printed fully soft robotics with highly flexible three-dimensional locomotion at the liquid–air interface. ACS Nano 16, 19393–19402 (2022). [DOI] [PubMed] [Google Scholar]
  • 27.Pang X., Lv J.-A., Zhu C., Qin L., Yu Y., Photodeformable azobenzene-containing liquid crystal polymers and soft actuators. Adv. Mater. 31, e1904224 (2019). [DOI] [PubMed] [Google Scholar]
  • 28.Zhang Y., Ben Jar P.-Y., Shifeng X., Li L., Quantification of strain-induced damage in semi-crystalline polymers: A review. J. Mater. Sci. 54, 62–82 (2019). [Google Scholar]
  • 29.Zhang H., Wu Y., Yang J., Wang D., Yu P., Lai C. T., Shi A.-C., Wang J., Cui S., Xiang J., Zhao N., Xu J., Superstretchable dynamic polymer networks. Adv. Mater. 31, e1904029 (2019). [DOI] [PubMed] [Google Scholar]
  • 30.Hermans P. H., Platzek P., Beiträge zur Kenntnis des deformationsmechanismus und der feinstruktur der hydratzellulose. Kolloid Z. 88, 68–72 (1939). [Google Scholar]
  • 31.Jiang Z., Tang Y., Rieger J., Enderle H.-F., Lilge D., Roth S. V., Gehrke R., Wu Z., Li Z., Men Y., Structural evolution of tensile deformed high-density polyethylene at elevated temperatures: Scanning synchrotron small- and wide-angle X-ray scattering studies. Polymer 50, 4101–4111 (2009). [Google Scholar]
  • 32.Bagheri M., Pourmoazzen Z., Synthesis and properties of new liquid crystalline polyurethanes containing mesogenic side chain. React. Funct. Polym. 68, 507–518 (2008). [Google Scholar]
  • 33.Zhang L., Liu Z., Wu X., Guan Q., Chen S., Sun L., Guo Y., Wang S., Song J., Jeffries E. M., He C., Qing F. L., Bao X., You Z., A highly efficient self-healing elastomer with unprecedented mechanical properties. Adv. Mater. 31, e1901402 (2019). [DOI] [PubMed] [Google Scholar]
  • 34.Zhang L., You Z., Dynamic oxime-urethane bonds, a versatile unit of high performance self-healing polymers for diverse applications. Chin. J. Polym. Sci. 39, 1281–1291 (2021). [Google Scholar]
  • 35.Zhao Q., Zou W., Luo Y., Xie T., Shape memory polymer network with thermally distinct elasticity and plasticity. Sci. Adv. 2, e1501297 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Text

Figs. S1 to S26

Tables S1 to S3

Legends for movies S1 to S4

References

sciadv.adk5177_sm.pdf (2MB, pdf)

Movies S1 to S4

sciadv.adk5177_movies_s1_to_s4.zip (54.1MB, zip)

Articles from Science Advances are provided here courtesy of American Association for the Advancement of Science

RESOURCES